[PATCH] Fix bug #13121 - Non-smbd processes using kernel oplocks can hang smbd

[Jeremy Allison]

On Mon, Nov 27, 2017 at 07:14:05PM +1300, Andrew Bartlett wrote:
> 
> From the build output I've seen It isn't looking good so far.  We have
> this test fail at least as often as it succeeds on cloud VMs.  Would it
> help if we gave you access to such a VM to work on?

This only seems to be happening on your cloud VM's, not on
the regular build machines. I also can't reproduce locally.

It's showing a missing RT signal from the kernel to the
lease holding process. I can't currently see any errors
in the test code for this (and indeed any tests I do always
show the correct RT signal being delivered to the correct
lease-holding process).

I will try and track this down, but it seems to be specific
to your environment so until you can show it happening elsewhere,
this isn't going to be a high priority for me.